Dear TF Hub Team,

USE paper Section 5 has a interesting paragraph on evaluation where authors use Arc Cosine (Cos Inverse) whose range is 0 to Pi in radians instead of plain cosine distance with range 0 to 2.

". For the pairwise semantic similarity task, we directly assess
the similarity of the sentence embedding produced by our two encoders. As show

I tried building the docs, but was met with a graphviz error. Typically this means I can spend a few hours pecking away at the dependencies until I get stable build... or someone that has it working can export their environment, and publish an environment.yml that we can use with the build instructions.
I was going off of the d2l book since that's a dep here, but their [environment.yml](https://g

I'm playing around with this wonderful code but I'm running into a curious issue when I try to train the model with my own data.

I replicated the personachat_self_original.json file structure and added my own data. I deleted dataset_cache_OpenAIGPTTokenizer file but when I try to train, I get this error:

INFO:train.py:Pad inputs and convert to Tensor
Traceback (most recent call last)

Hey! I think it would be useful to have a more detailed explanation about:

1. what the dataset should look like for performing NER, similar to the fine-tuning example. The [NER sample](https://github.com/deepset-ai/FARM/blob/97b0211a37ea7c7d64b4602f0e21b65428b2bd76/t

Hi,

I noticed that I cannot use nlp.update with Doc as inputs instead of raw strings. Please refer to the mini example below:

from spacy_transformers.language import TransformersLanguage
from spacy_transformers.pipeline.wordpiecer import TransformersWordPiecer
from spacy_transformers.pipeline.tok2vec import TransformersTok2Vec

nlp = TransformersLanguage(trf_name='test_bert', meta

Per this comment in #12